aaolcay / autoencoders

Autoencoders are mostly used for different purposes such as denoising, compression data, anomaly detection, generating new data from the input data entering to the model, and more. This repository introduces a simple autoencoder architecture with some brief explanations of encoder, bottleneck and decoder parts.

Autoencoders

In this repository, different autoencoder applications are published with introductions for readers or researchers to understand the major concepts. The good thing, here on this repository, is that everything was made easy for someone who just started studying DLNNs. We would insert all things into the bunch of class but thought that it might have made some troublesome for comprehension. Therefore, we just gave all the things in our codes line by line.

We updated the python code that you can see in 'autoencoder_updated.ipynb' so that we can include how to save and load a model built for using similar tasks in different problems (i.e., transfer learning). We also compared both original data and autoencoded data.

Latest update: 25/02/2023